When we chose Learning for the Jobs of Today, Tomorrow, and Beyond as our Intersect 2017 conference theme, we were thinking about the journey from learning to a job at a fairly high level, as we wanted to embrace something that would resonate through every aspect of the conference.
Once the planning stages were largely behind us, we were able to really start looking at the meaning of this theme at eye-level, and thinking about what it means to each individual student.
We found ourselves in particular thinking about our Self-Driving Car Engineer Nanodegree program students, as in many ways they are preparing themselves for an industry that is still in the imagination stages! Self-driving cars are coming. We know this, as do our students. But the roles they’re now preparing for truly are the jobs of tomorrow.
Artificial intelligence. Machine learning. Self-driving cars. If you’re keeping up with the rapid changes in the technology industry, you’re seeing a bunch of terms thrown around as if they’re interchangeable—but really, there are some pretty important distinctions. In this post, we’re going to demystify the differences, and clarify the relationships, among these terms, especially artificial intelligence, machine learning, and self-driving cars. Let’s begin with a simple model for how we’ll approach this topic:
The energy around our Self-Driving Car Engineer Nanodegree program right now is incredible. It seems like only yesterday that Sebastian Thrun announced the program at TechCrunch Disrupt SF. Since then, we’ve reviewed and accepted applications, and confirmed the enrollments of our first groundbreaking students. We’ve opened a Slack channel that is now buzzing with thousands of conversations. We’ve even started conducting challenges for people to begin contributing to our very own autonomous vehicle!
But today’s news takes things to a whole new level of amazing.
Launching something like our Self-Driving Car Engineer Nanodegree program is, to put it mildly, complicated. The technology, the logistics, the curriculum, the applications, the partner relationships, the structure, the finances, the messaging, the marketing, the legal implications, the support requirements, not to mention the self-driving car itself—it all combines to make the endeavor quite an undertaking! The effort is company-wide, and the immersion full-bore.
But there is more to it than just getting the program up and running.
As this whole crazy thing was transforming from idea into reality, there was an extraordinary amount of soul-searching taking place. Every single person at Udacity was thinking through what it all meant, in order to try and ensure we were doing the right thing, and truly creating something incredible for our students. Some of this took place at a personal level, and much of it transpired in small team meetings, as different groups within the company worked through their respective challenges to achieve their goals.
At certain points however, these smaller-scale self-queries became universal soul-searches, and there would be moments when every single person at Udacity had to stop and wonder, are we really doing the right thing?
Just such a moment occurred recently, when an employee posted the following question to an internal forum:
Could someone explain in more detail why we’re building a self-driving car?
I guess I just don’t understand too well what this project has to do with democratizing education. Could someone go over what the company is hoping to achieve by doing this, in terms of our values and ultimate goals as a company and a force for good in the world?
I know that when I first read this, I kind of froze inside for a moment, suddenly unsure of whether we’d somehow lost our way along the way. But then, the responses began to come in.
Because without a real car to test on, our students won’t actually be self-driving car engineers.
In order to build real competence, and for our program to provide any credibility to our graduates, they need the opportunity to work with a real self-driving car. Our Virtual Reality students can use Cardboard, our Predictive Analytics students can use Alteryx and Tableau, our Machine Learning students can build and test real models against real data, our developer students can build real web apps… But without a real car to work with, Self-Driving Car students’ educations are crippled.
And as to why we’re building one rather than buying one—what better way to build real, deep experience and expertise than to build something from the ground up?
This is just one example of the kinds of responses that were posted. But as eloquent and reasoned as that comment is, I think my personal favorite was probably this one (what a great headline!):
Because it’s so freakin awesome?
For me, looking at it from a student perspective, I would never dream of being able to actually work on a self-driving car as part of my education. It’s the ultimate experience, getting the knowledge plus being able to implement that knowledge which is one of the things we’ve been trying to do for our students. Giving them the platform to implement their knowledge in projects, and the ultimate project here for a self-driving car engineer, is to actually program a self-driving car!
I don’t know that we could ever turn that into a marketing headline, but just once, it would be kind of fun to run a campaign like that! Enroll today in our Self-Driving Car Engineer Nanodegree program, because it’s so freakin’ awesome!
One of the questions that has of course occupied us throughout the process of bringing this program to life is cost, and one of the forum responses addressed this issue head on:
My $0.02, coming from a robotics background: robotics is absurdly expensive. Being able to centralize the hardware and democratize access to it greatly reduces the costs and barriers to entry for aspiring roboticists and self-driving car engineers, alike. Isn’t that democratizing education at its heart? Providing access and enfranchising aspiring engineers who normally wouldn’t have access?
Random aside: check out the fees for FIRST Robotics: $5-6k, not including parts, travel, supplies, etc. That’s for a high-school robotics team. We’ll be able to deliver access to a $100k+ self-driving car for a fraction of that cost.
Having read this far, I was feeling pretty overjoyed. I felt total confidence in the program, and absolute adoration and admiration for my colleagues. So much so, that when I moved on to the following, the succinctness of this reply felt like a perfect summary:
Students will be able to run their code on our car. This creates an amazing opportunity for students, and gives access to technology they might not otherwise gain access to.
I love being part of an organization that takes up these issues internally, and works through them communally. I love that people ask questions, challenge assumptions, and ask for explanations. I love that people care enough to answer, and maybe best of all, I love that we appreciate one another for doing all the above. In this particular case, I especially love what the individual who first asked the question had to say in response:
Ok, see this is why I posted the question! Thank you—I understand it now a lot more.
Please know, I don’t write any of the above to try and prove how wonderful Udacity is, or to try and affirm that we’ve gotten everything right with this program. I do think Udacity is made up of passionate, caring, talented, and thoughtful individuals, and I do think the Self-Driving Car Engineer Nanodegree program is a pretty incredible offering, but the real reason I want to share this, is to try and provide a small glimpse into the kinds of self-queries we put ourselves through as we work to create new opportunities for you, our students.
It all comes down to the “Students First” ethos. That’s the idea we come back to every time we make a decision, launch an offering, engage in an experiment, contemplate a change, or do anything that will impact our students. Have we put our students first? That’s the question. It is always the question. And we will always try and make sure we can answer “yes.”
In 2004, DARPA held the first Grand Challenge for an autonomous car to drive 142 miles through the Mojave desert in under 10 hours. Fifteen cars participated, none finished. In 2005, the Challenge was repeated and 23 cars entered. Four finished under 10 hours and our car “Stanley” won in 6 hours and 53 minutes—11 minutes ahead of the next car. Then, in 2010, the Google self-driving car navigated 1,000 miles of public roads in California, an unbelievable advance from that first challenge.
Today, six years later, autonomous cars have become one of the hottest areas for innovation. The Boston Consulting Group estimates the market for autonomous cars will hit $42 billion in 2025. The World Health Organization reports there are 1.2 million traffic fatalities globally every year and driverless cars could be poised to save no less than 1 million lives per year.
Technology companies, automotive manufacturers, media giants, and start-ups around the world are rapidly pushing new advances in this space, whether it be hardware or software. And, they all need talent.
This is where Udacity comes in. Today, we are announcing that our Udacity Self-Driving Car Engineer Nanodegree program is open for students to apply. It is the first and only program of its kind where most people with an internet connection—from Detroit to Damascus and from Adelaide to Aleppo—can learn the skills they need to work in one of the most amazing fields of our time.